1) Structure/function studies of the HIV Env trimer In collaboration with Dr. Paolo Lusso, NIAID, we have obtained strong experimental evidence in favor of a novel model whereby gp120 adopts a specific closed conformation in the native Env trimer. The evidence is based on studies with synthetic peptides, engineered gp120 variants, monoclonal antibodies, molecular dynamics simulations, and fitting analyses with published trimer structures. In addition to new insights into Env structure and its mode of protection from neutralizing antibodies, these results also suggest potential new paths for vaccine design. 2) We have initiated studies to design protein constructs displaying the extracellular regions of the HIV-1 coreceptors CCR5 and CXCR4, to facilitate biochemical and structural studies of the gp120-coreceptor interactions. Various designs are being pursued, including the linking of these domains to stable protein scaffolds. 3) Immunotoxiins for targeted killing of HIV-infected cells. We have collaborated in the use of in vivo models to test the ability of immunotoxins directed against HIV-1 Env expressed on the surface of productively infected cells to deplete infected cell reservoirs persisting during ART. In collaboration with Dr. Tom North (U. C. Davis/Emory U.), we examined the effects of sCD4-PE40 in rhesus RT-SHIV-infected macaques whose viral loads have been suppressed below detectable levels by combination ART. The results indicated that augmentation of ART with immunotoxin significantly depleted persisting viral RNA compared to ART alone in gut tissues (1-3 log10 in jejunum, ileum and cecum); enhanced depletion was also observed in other tissues, though to lesser extent. Clearly under the conditions employed, the virus was not eliminated; viral rebound after cessation of all treatment was similar in animals treated with ART alone compared to ART plus immunotoxin. In a collaboration with Dr. Victor Garcia-Martinez (U. North Carolina) to test the activity of CD4-PE40 and 3B3-PE38 in HIV-1-infected humanized mice (BLT mice), we found that compared to ART alone, ART plus the 3B3-PE38 immunotoxin profoundly depleted infected cells systemically by nearly 10-fold; particularly dramatic effects occurred in specific tissues/organs (0.4 - 3 log10 reductions). However the virus was not eradicated; therefore it is likely that if experiments are repeated and the animals not sacrificed for analysis of tissues, viral loads will rebound after cessation of all therapy. 4) Novel chimeric antigen receptors (CARs) for adoptive transfer of autologous CD8 T cells genetically modified to target HIV-1 Env. The immunotoxin approach suffers from the inherent problem of immunogenicity due to the foreign moiety of such proteins (bacterial in the case of PE-based agents); thus the immunotoxin treatment is limited to short duration, a major limitation in view of the presence of reservoirs of latently infected cells that can become activated at a later time. Thus a more durable targeted cell killing method would seem preferable. Adoptive transfer of engineered CD8 T cells represents an exciting option, particularly in view of the profound successes of this strategy in the cancer field. This approach was tried for HIV infection by several groups more than a decade ago; while the transferred engineered T cells were found to populate many anatomical sites where HIV replication occurs, they persisted only at low levels, and no clinical benefits were obtained. In collaboration with the group of Dr. Steven Rosenberg, NCI, we have introduced two new features to the CAR constructs that are designed to improve efficacy of this approach. The first includes the use of 2nd and 3rd generation intracellular signaling domains developed by cancer investigators in the CAR field namely regions of CD28 and 41BB; these are expected to enhance the functional activity and in vivo longevity of the adoptively transferred CD8 cells. Perhaps more importantly are modifications that we have introduced into the extracellular Env-targeting moiety. We have designed modified variants of the 2-domain extracellualr region of CD4 (designated CD4-M). In vitro experiments have revealed two major improvements of CD4M compared to CD4 as the Env-targeting moiety. First, the CD4M CARs are devoid of the unwanted activity of CD4 CARs in rendering transduced CD8 cells susceptible to HIV-1 infection (via CCR5 endogenously expressed on these cells). Second, we have made several CD4M-CAR variants, one of which shows superior anti-HIV-1 efficacy compared to the previously used CD4-CAR. Collaboration is underway with Mario Roederer, NIAID VRC, to test the various CARs in rhesus macaques infected with SHIV or SIV viruses. Also planned are collaborative studies to test the CARs in murine models.